Several mechanical models with lumped mass were employed in the literature to represent strongly frequency-dependent characteristics of hydraulic engine mounts. Although complex stiffness by the mechanical models showed good agreements with the measured values, there exists a critical pitfall. The fact is that the complex stiffness model was derived at the driving point while measurements were obtained across the mounts. That is, the point stiffness model was mistakenly compared with the transfer stiffness measurement. In this paper, limitations of the mechanical model with lumped mass are discussed by illustrating actual measurement results of the stiffness matrix.